Seasonality in surveys : A comparison of Belgian, French and German business tests

Seasonal Effects in Qualitative Responses of Monthly Business Surveys, Modeled with Log-Linear Probability Methods, Can Be Extracted Like in Linear Regression Or Some Time Series Models by Using Dummy Variables. There Is Another Route to Extract the Seasonal Effect, Which Is to Unique to Surveys, by Incorporating Questions on the Questionnaires Which Force Respondents to Make a Distinction Between Seasonal and Cyclical Effects. the Belgian, French and German Surveys Allowed Us to Compare the Alternative Approaches to Seasonality. We Estimated a Model of Micro-Based Firm Behavior and Found the Approach with Survey-Based Seasonal Measures More Desirable. Dummy Variables Pick Up Aggregate Seasonal Effects Across Industries. the Advantage of Survey-Based Measures of Seasonal Effects Is That They Are Firm-Specific Observations Instead of Aggregate and Hence Tend to More Precise. Moreover, the Responses to Requests to Separate Seasonal From Cyclical Effects Seems to Be Fairly Consistent.